AI scribes—tools that transcribe and draft clinical documentation—are being adopted rapidly across healthcare settings. Alongside adoption, multiple concerns have been raised regarding accuracy, transparency, consent, and legal responsibility.
Key issues discussed in recent reporting include:
- Documentation accuracy and error rates: Studies have reported error rates in the range of 1–3%, including omitted details, misinterpretations, and fabricated content (often referred to as “hallucinations”).
- Consent and transparency: A lawsuit involving Sharp HealthCare alleges a patient encounter was recorded by an AI scribe (Abridge) without the patient’s knowledge or consent. The patient further alleges the resulting documentation stated consent had been obtained.
- Liability and legal exposure: It remains unclear how responsibility will be allocated when AI-generated documentation contains clinically meaningful errors. Persistent audio recording and storage may also introduce additional legal considerations for health systems.